Brazilian Journal

of Chemical ISSN 0104-6632

Printed in Brazil
Engineering www.abeq.org.br/bjche

Vol. 30, No. 03, pp. 551 - 562, July - September, 2013

MASS TRANSFER OF SO2 ABSORPTION WITH

AN INSTANTANEOUS CHEMICAL REACTION
IN A BUBBLE COLUMN
Xiaolei Li1, Chunying Zhu1*, Sumin Lu2 and Youguang Ma1
1
School of Chemical Engineering and Technology, State Key Laboratory of Chemical Engineering,
Phone: +86 (0) 22 2740 4772, Fax: +86 (0) 22 2740 4772, Tianjin University, Tianjin, 300072, China.
E-mail: zhchy971@tju.edu.cn
2
Department of Material and Chemical, Tianjin Polytechnic University, Tianjin, 300160, China.

(Submitted: April 22, 2012 ; Revised: June 18, 2012 ; Accepted: August 3, 2012)

Abstract - Gas absorption accompanied by an instantaneous irreversible chemical reaction in bubble columns
has been analyzed theoretically. A mass transfer model based on the Dankwerts surface-renewal model as
well as the penetration theory for surface stretch proposed by Angelo et al. was developed, in which the
effects of bulk motion and turbulence on mass transfer were taken into account. The analytical expressions for
the time-average mass transfer coefficient and the enhancement factor have been obtained. The fast reactive
absorption of SO2 from gas mixtures into aqueous NH4HCO3 solution was investigated experimentally in a
bubble column reactor to validate the mass transfer model, and the results calculated by the present model
agree well with the experimental results.
Keywords: Absorption; Bubble column; Mass transfer; Instantaneous reaction; Enhancement factor.

INTRODUCTION hydrogen sulphide by means of sodium hydroxide

Gas absorption accompanied by chemical
reactions is widely encountered in chemical and
allied industries. A number of experimental and
theoretical studies on this type of gas absorption
process have been reported over the past few decades
and many mass transfer models, as well as their
corresponding numerical or analytical solutions,
have also been developed based on the film theory,
the penetration theory or the surface-renewal model
(Danckwerts, 1950; Last et al., 2002; da Silva et al.,
2007; Danish et al., 2008; Meldon et al., 2007;
Kakaraniya et al., 2007; Jun Yue et al., 2012; Zhang
et al., 2012; Martínez et al., 2012). A chemical
reaction can be considered to be instantaneous when
its rate is much greater than the rate of diffusion.
Some typical examples include the removal of
solution and the absorption of ammonia into sulfuric-
acid solution, etc. In fact, many chemical absorption
processes can be considered to be instantaneous
reactions, e.g., the removal of acidic gases with
alkaline solutions, the removal of alkaline gases with
acid solutions, etc. For gas absorption accompanied
by instantaneous chemical reactions, the solute
diffuses from the gas phase into the liquid phase and
reacts immediately and completely with the reactant
present in the solution, and then a sharp reaction
plane parallel to the gas-liquid interface is formed.
Danckwerts (1950) developed a mathematical model
to describe mass transfer and heat-conduction with a
moving boundary for the process of gas absorption
accompanied by an instantaneous chemical reaction
in an unsteady-state using the film theory. Garg et al.
(2000) focused on the moving-boundary problem of

*To whom correspondence should be addressed

552 Xiaolei Li, Chunying Zhu, Sumin Lu and Youguang Ma
gas absorption accompanied by an instantaneous
chemical reaction in a limited gas-liquid system. Jun
Yue et al. (2012) also studied gas absorption with
instantaneous reaction in a finite liquid layer.
However, because the conventional model for gas
reactive absorption generally introduced empirical
parameters to reflect the effects of the bulk motion
and turbulence, it does not reveal the inherent
mechanism of the influence of the bulk motion and
turbulence on mass transfer between the gas-liquid
phases, especially in a bubble column with very
complex two-phase flow behavior. Inspired by this
point, Angelo et al. (1966) developed the penetration
concept in the absence of bulk motion and turbulence
and proposed a promising theory called “surface-
stretch theory”. Jajuee et al. (2006) further developed
the model by incorporating Dankwerts (1951)
surface-renewal model with the penetration theory
for surface stretch proposed by Angelo et al (1966)
and extended the model to liquid-liquid systems.
In this study, a theoretical analysis of gas absorp-
tion accompanied by an instantaneous chemical
reaction of the type “A + YB = products” is presented
taking into account the effects of bulk motion and
turbulence. A mass transfer model based on the
Dankwerts (1951) surface-renewal model and the
penetration theory for surface stretch proposed by
Angelo et al. (1966) was developed and analytical
solutions have been derived. The mass transfer process
of the fast reactive absorption of SO2 into aqueous
NH4HCO3 solution in a bubble column was experi-
mentally investigated to validate the proposed mass
transfer model.
Figure 1: Schematic diagram of the experimental setup
Brazilian Journal of Chemical Engineering
EXPERIMENTAL
A schematic diagram of the experimental setup,
which consisted of the mixing gas generation unit,
exhaust detection devices and bubble column, is
shown in Figure 1. The bubble column was made of
stainless steel and was 20 cm in diameter and 45 cm
in height.
SO2 with a volume concentration of 98% (Tianjin
Kermel Chemical Reagent Co., Ltd) supplied from a
cylinder was diluted with air from the air compressor
(Shanghai Jiebao Compressor Manufacture Co., Ltd.).
The two kinds of gases were mixed in the gas mixer.
Then the air-SO2 mixture was fed at the bottom of
the column and the SO2 reacted with NH4HCO3
(mass purity > 0.995, Tianjin Kermel Chemical Reagent
Co., Ltd) immediately in the liquid phase. Air was
fed into the bubble column before the experiments.
When the system reached stable conditions, valves
(3) and (4) on the SO2 cylinder were opened and the
air-SO2 mixture introduced into the bubble column.
The concentration of SO2 in the exhaust was detected
by the flue gas analyzer (Qingdao Minhope
electronic instrument Co., Ltd). The liquid was
sampled at the bottom of the column in the SO2
absorption process in order to measure the pH with a
digital pH meter. The experiments were carried out
at room temperature and repeated three times. The
gas flow rates were controlled by rotameters (LZB-
type, Tianjin flow Instrument Co., Ltd.) with the
measuring range of 0-100 ml·min-1 for SO2 and
0-100 L·min-1 for air; the accuracy of rotameters was
±1.5%.
1) Sulfur dioxide cylinder;
2) Air Compressor;
3) Valve;
4) Pressure reducing valve;
5) Manometer;
6) Rotor flow meter;
7) Gas mixer;
8) Bubble column;
9) Feed tank;
10) Flue gas detector;
11) Manometer.
 Mass Transfer of SO2 Absorption with an Instantaneous Chemical Reaction in a Bubble Column 553

THEORIES A( θ) = − A ∞ (1 − e −sθ ) + A 0 (6)

Mass Transfer Model Consider a small flat element of surface ΔxΔy

For gas absorption accompanied by an instantane- moving with the fluid bulk velocity, the change in
ous chemical reaction of the type A + YB = products, the area of the surface of this flat element with time is:
the heat of reaction could be neglected because of
the excellent heat transfer characteristics in the ⎡ Δx(u y y +Δy − uy y ) ⎤
bubble columns when the concentration of the gas- ∂A ⎢ ⎥
( )= (7)
∂θ ⎢ ⎥
phase reactant is low. It is assumed that diffusion ⎢⎣ +Δy(u x x +Δx − u x x ) ⎥
⎦
occurs only in the direction perpendicular to the
interface. The model describing the process can be
stated as follows: Dividing by ΔxΔy and taking the limit as Δx and
Δy approach zero, Eq. (7) may be rewritten as:
∂c A ∂c ∂ 2c A
− u z A = DA for 0 ≤ z ≤ z λ (1)
∂θ ∂z ∂z 2 ∂ ln A ∂u ∂u y
( )= x + (8)
∂θ ∂x ∂y
IC. θ = 0 ⇒ c A = 0 for all z
For an incompressible fluid, Eq. (8) can take the
BC. z = 0 ⇒ c A = A* , z = λ ⇒ c A = 0 , for θ > 0 following form with the aid of the equation of
continuity:
∂c B ∂c ∂ 2c
− u z B = D B 2B for z λ ≤ z ≤ σ (2)
∂θ ∂z ∂z ∂ ln A ∂u
( )=− z (9)
∂θ ∂z
IC. θ = 0 ⇒ c B = B* for all z
Since the expression on the left-hand side is
BC. z = λ ⇒ c B = 0 , z = σ ⇒ c B = B* , for θ > 0 independent of z, it may be obtained for negligible
net transfer across the interface:
In the position of the reaction plane, the following
equation can be obtained: ∂ ln A
u z = − z( ) (10)
∂θ
dc A dc B
YD A z =λ = − DB z =λ (3)
dz dz To solve Eqs. (1) and (2), we define the following
dimensionless variables:
According to the surface-renewal model of
Danckwerts (1951), the surface-age distribution
Z1 = z D A θ0 (11)
function is given by φ( θ) :

φ( θ) = Se −Sθ (4) Z2 = z D Bθ0 (12)

The film area comprising the elements having T = θ θ0 (13)

ages between θ and ( θ + dθ) is φ( θ)dθ . Let the total
area of the surface exposed to the gas be equal to
A( ∞) and A( θ) for the surface areas with ages C A = c A A* (14)
between θ and ( θ + dθ) . One can then obtain:
C B = c B B* (15)
dA( θ)
= φ( θ)dθ = Se −sθdθ = −d(e −sθ ) (5)
A( ∞) where θ0 is a constant that can be evaluated from the
time characteristics of bubble column, the bubble-
Eq. (5) may be readily integrated and taking A = A 0 formation time. Using the dimensionless variables,
at θ = 0 gives: Eqs. (1) and (2) take the following forms:

Brazilian Journal of Chemical Engineering Vol. 30, No. 03, pp. 551 - 562, July - September, 2013
554 Xiaolei Li, Chunying Zhu, Sumin Lu and Youguang Ma

∂C A ∂ ln A ∂C ∂ 2C A ⎡ ∂ ln A ⎤
− Z( ) x,y A = D A (16) ⎢ ( ∂T ) x,y G ' ⎥
∂T ∂T ∂Z1 ∂Z12 ⎢ − 3⎥ = α (22)
⎢ G2 G ⎥
IC. T = 0 ⇒ CA = 0 for all Z ⎣ ⎦

where α is an arbitrary constant. The following

BC. Z1 = 0 ⇒ CA = 1 , assumption was also made:

Z1 = λ1 = λ D A θ0 ⇒ C Α = 0 ∂ ln A
( ) x,y = F (23)
∂T
∂C B ∂ ln A ∂C ∂ 2C B
− Z( ) x,y B = D B (17) Eq. (22) may be rearranged to:
∂T ∂T ∂Z2 ∂Z22
FG − G ' = αG 3 (24)
IC. T = 0 ⇒ C B = 1 for all Z
Eq. (24) is a Bernoulli equation, which can be put
'
BC. Z2 = λ 1 = λ D B θ0 ⇒ C B = 0 , into linear form by the following substitution:

G = h −1 2 (25)
Z2 = σ1 = σ D B θ0 ⇒ C B = 1
By using Eq. (25), Eq. (24) takes the form:
In order to solve these equations, combinations of
the variables are introduced to convert the partial dh
differential Eqs. (16) and (17) to simple differential = −2Fh + 2α (26)
dT
equations:
The integrating factor can be introduced:
C A = CA ( φ1 ) C B = C B ( φ2 ) (18)

e 2 ∫ FdT = e
(
2 ∫ ∂ ln A
∂T )dT = A(T)2 (27)
where

φ1 = Z1 × G ( T ) φ2 = Z2 × G ( T ) (19) Combining Eqs (26) and (27), one obtains:

Eqs. (16) and (17) thus become: A2

dh
dT
= −2hA
dA
dT
+ 2αA 2 ⇒ d A 2 h = 2αA ( ) (28)

⎡ ∂ ln A ⎤
d 2C A ⎢ ( ∂T ) x,y G ' ⎥ dCA Eq. (28) can be integrated:
+ φ1 ⎢ − 3⎥ =0
dφ12 ⎢ G2 G ⎥ dφ1 T
A 2 h = 2α ∫0 A(T) dT + C5
2
⎣ ⎦ (29)
(20)
for 0 ≤ φ1 ≤ φ1,λ
Eq. (29) in transformed form becomes:

⎡ ∂ ln A ⎤ T
d 2C B ⎢ ( ∂T ) x,y G ' ⎥ dC B
+ φ2 ⎢ − 3⎥ =0
G (T ) = A (T ) 2α ∫0
A ( T ) dT + C5 (30)
dφ22 ⎢ G2 G ⎥ dφ2
⎣ ⎦ where C5 is a constant.
(21)
for φ2,λ ≤ φ2 ≤ φ2,σ Then the solutions of Eqs. (20) and (21) are
(Mehra, 1990):
where G ' indicates the partial derivative of G with
respect to T at constant Z. It has been assumed that: CA = C1 + C2erf ( 2 αφ1 2 ) (31)

Brazilian Journal of Chemical Engineering

 Mass Transfer of SO2 Absorption with an Instantaneous Chemical Reaction in a Bubble Column 555

C B = C3 + C4erf ( 2 αφ2 2 ) (32) Y D A A*

e
−φ2
1,λ ⎡erf ( φ2,δ ) − erf ( φ2,λ )⎤
D B B* ⎣ ⎦
where C1, C2, C3 and C4 are constants. If the assumed (39)
−φ2
combinations of variables are successful, equations =e 2,λ
erf ( φ1,λ )
(23) and (24) should satisfy the initial and boundary
conditions of Eqs. (16) and (17).
The instantaneous rate of gas absorption for a
CA ( 0 ) = 1 CA ( φ1,λ ) = 0 (33) surface element may be defined by:

dA
C B ( φ2,λ ) = 0 C B ( φ2,σ ) = 1 (34) dN A = jA,z = jA,z φ( θ)dθ (40)
A∞

According to the literature (Jajuee et al., 2006), where

for convenience α=2 is taken. Thus, C5 in Eq. (30)
should be zero to satisfy Eq. (33). Combining Eqs. ∂c A ∂CA ∂Z
(31) and (33) yields: jA,z = − D A z =0 = D A A* Z=0 (41)
∂z ∂Z ∂z

CA = 1 − erf ( φ1 ) erf ( φ1,λ ) (35) Hence

Combining Eqs. (32) and (34) leads to the following 1 2

dN A = D A A*
relation: erf ( φ1,λ ) π
(42)
erf ( φ2,λ ) ⋅ G (T ) ⋅
1
⋅ Se −Sθ ⋅ dθ
CB = −
erf ( φ2,δ ) − erf ( φ2,λ ) D A θ0
(36)
erf ( φ2 ) Eq. (42) can be integrated to get the average flux
+
erf ( φ2,δ ) − erf ( φ2,λ ) over the time of exposure per unit area of turbulent
surface:
From Eq. (35) one gets:
∞ 1 2 S2 D A
∂c A ∂C ∂C ∂Z
N A = A* ∫0 erf ( φ1,λ ) πθ0
⋅ G ( T ) ⋅ e − Sθ ⋅ d θ
= A* A = A* A =
∂z ∂z ∂Z ∂z (43)

1 2 −φ12 G ( T )
(37) = KL A − 0 ( *
)
− A* ⋅ e ⋅
erf ( φ1,λ ) π D A θ0
The time-average mass transfer coefficient can be
obtained:
From Eq. (36) the following equation can be
obtained:
∞ 1 2 S2 D A
∂c B ∂C ∂C ∂Z
= B* B = B* B =
KL = ∫
0 erf ( φ1,λ ) πθ0
⋅ G ( T ) ⋅ e −Sθ ⋅ dθ (44)

∂z ∂z ∂Z ∂z
(38)
de Lind van Wijngaarden et al. (1986) pointed
1 2 −φ22 G (T )
B* e ⋅ out that the “equal diffusivities” condition was a
erf ( φ2,δ ) − erf ( φ2,λ ) π D B θ0 severe limitation. Garg et al. (2000) derived analytical
solutions for the case of equal diffusivities for gas
At the position of the reaction plane, λ, absorption accompanied by an instantaneous chemi-
substituting Eqs. (37) and (38) leads to the function cal reaction in finite gas-liquid systems. The “equal
for the change of the position of the reaction with diffusivities” condition is used at z = λ , the position
time: of reaction plane. From Eq. (39) at z = λ :

Brazilian Journal of Chemical Engineering Vol. 30, No. 03, pp. 551 - 562, July - September, 2013
556 Xiaolei Li, Chunying Zhu, Sumin Lu and Youguang Ma

erf ( φ1,λ ) = erf ( φ2,λ )

( )
14
(45) l = k3 ε (50)

The function for the variation of the position of

υ = ( kε )
14
the reaction plane with time may be obtained: (51)

Y D A A*e −1 DA D B B*e −1 D B where k is the kinematic viscosity:

erf ( φ1,λ ) =
1 + Y D A A*e −1 DA D B B*e −1 D B
(46) k =μ ρ (52)
erf ( φ2,δ )
ε is the viscous dissipation of energy per unit mass of
eddies, which can be calculated from (Wang et al.,
By substituting Eq. (46) into Eq. (44), the mass 2005):
transfer coefficient is found to be:
⎛ εg ⎞
Y D A A*e −1 DA ε = ⎜ Ug − UL ⎟ g (53)
1+ ⎜ 1 − εg ⎟
D B B*e −1 D B ⎝ ⎠
KL =
Y D A A*e −1 DA
where Ug is the superficial gas velocity, UL the
D B B*e −1 D B superficial liquid velocity and εg the gas hold-up. For
(47) bubble columns with low superficial liquid velocity,
D AS ∞ 1 2 S UL can be regarded as zero and the energy dissipation
π ∫0
⋅
erf ( φ2,δ ) θ0
G ( T ) ⋅ e −Sθ ⋅ dθ
rate can be obtained:

ε = Ugg (54)
The expression for the mass transfer coefficient is
too complicated to be employed for practical
purposes due to the error function in the equation. The fractional rate of surface renewal may be
But when the scale of turbulence is high, Jajuee et al. defined as follows (Jajuee et al., 2006):
(2006) found that the integrand in Eq. (47) had the
value of 2. The scale of turbulence in bubble s=υ l (55)
columns is high; thus the expression of mass transfer
coefficient becomes: For mass transfer accompanied by chemical
reaction, one of the most important parameters is the
1 + Y D A A*e −1 DA D B B*e −1 D B enhancement factor, defined as the ratio of the time-
KL = mean mass flux at the interface with chemical
Y D A A*e −1 DA D B B*e −1 D B
reaction to the time-mean mass flux at the interface
(48) without chemical reaction under the same driving
4D AS force.
π
E = K L K*L (56)
Harriott (1962) pointed out that the concept of
surface-renewal theory could only be expected when The time-mean mass flux at the interface without
chemical reaction is (Jajuee et al., 2006):
H DS < 0.5 is met, where H is the average
distance to which eddies approach the interface.
K*L = 4D AS π (57)
Thus, the distance from the interface to the bulk of
the liquid is equal to H plus the length of eddies:
Substitution of Eqs. (48) and (57) into Eq. (56)
yields:
δ= H+l (49)
1
The length and velocity scales of eddies for E = 1+ * −1 D A
(58)
isotropic turbulence can be calculated from: Y DA A e D B B*e −1 DB

Brazilian Journal of Chemical Engineering

 Mass Transfer of SO2 Absorption with an Instantaneous Chemical Reaction in a Bubble Column 557

Bubble Column Model HSO3− = H + + SO32−

(65)
In this experiment, low superficial gas velocities K2 = C +C C
H SO32 − HSO3−
were applied, which was in the bubble flow regime.
The heat of reaction can be neglected due to the low
sulfur dioxide concentration and the excellent heat CO2 + H 2O = H + + HCO3−
transfer characteristics of bubble columns. Thus, the (66)
K3 = C +C C
mass balance of sulfur dioxide in the gas phase is: H CO32 − HCO3−

∂CSO2 ∂CSO2 ∂ 2CSO2 HCO3− = H + + CO32−

+ ub = Dg (67)
∂t ∂z ∂z 2
(59) K4 = C +C C
H CO32 − HCO3−
K L,tola
−
HεG
( HCSO 2
− RC L )
H 2O = H + + OH −
(68)
The mass balance of sulfur in the liquid phase is: Kw = C +C
H OH −

∂C L ∂ 2C L K L,tola
∂t
= DL
∂z 2
+
1 − εG
HCSO2 − RC L ( ) (60) NH 4+ + OH − = NH 3 + H 2O
(69)
K 6 = C NH 3 C − C
OH NH 4+
where the total mass transfer coefficients are given
by the following equation: The charge balance in the liquid is as follows:

1 1 1 C +C =C + 2C +C
NH 4+ H+ HSO3− SO32 − HCO3−
= + (61)
K L,tol *
EHK L k G (70)
+ 2C +C
CO32 − OH −

The gas-side volumetric mass transfer coefficient

can be calculated from (Cho et al. 1988): Combining Eqs. (63)-(65) leads to:

CSO2 ,L = RC L (71)
k G a = 2.6 × 103 DG
0.5 0.76
ub (62)
where
Before the NH4HCO3 is consumed completely,
CL in Eqs. (59) and (60) is zero and E is calculated
from Eq. (58). After the NH4HCO3 is consumed
R = C2 +
H (C 2
H+
+ K1C
H+
+ K1K 2 ) (72)
completely, CL is the total sulfur concentration in the
liquid: and E is calculated from (Bokotko et al., 2005):

C L = CSO2 ,L + C +C (63) E = 1+ D K1 DSO2 ( CSO2 ,L + CSO2 ) (73)

HSO3− SO32 − HSO3−

The following reactions in the bulk liquid should The chemical reaction equilibrium constants at
be considered: 25 °C are shown in Table 1. DSO2 D − and D −
HSO3 HCO3
are respectively 1.83×10-9 m2·s-1, 1.545×10-9 m2·s-1
SO 2 + H 2O = H 2SO3 = H + + HSO3− and 1.185×10-9 m2·s-1 at 25°C (Ebrahimi et al.,
(64) 2003). The Henry coefficient is 0.8082 atm·kg·mol-1
K1 = C +C CSO2
H HSO3− (Ebrahimi et al., 2003).

Table 1: The chemical reaction equilibrium constants

Parameter K1 K2 K3 K4 K5 K6
Value(mol·kg)-1 0.0144 7.0786×10-8 4.3766×10-7 4.6541×10-11 1.0038×10-14 6.5849×10-12

Brazilian Journal of Chemical Engineering Vol. 30, No. 03, pp. 551 - 562, July - September, 2013
558 Xiaolei Li, Chunying Zhu, Sumin Lu and Youguang Ma

Dispersion coefficient (Meikap et al., 2002), Effect of the Inlet SO2 Concentration
liquid phase:
Figure 2 shows the effect of the inlet SO2
D L = 0.678d1.5 0.3 1.5 0.3
R u b and d R u b < 400 (74) concentrations on the gas absorption. It can be found
from Fig. 2 that the duration time for complete SO2
Dispersion coefficient (Meikap et al., 2002), gas removal decreases with the increase of the inlet SO2
phase: concentration; in the last period of SO2 absorption,
the outlet SO2 concentration increases more
D g = 50d1.5 significantly with the increase of the inlet SO2
R u b εG (75)
concentration. The outlet SO2 concentration and the
pH value of the bulk liquid are plotted against time
Bubble rising velocity (Moo-Young et al., 1981):
in Figure 3. It can be seen that the pH value
decreased sharply from 6.17 to 3.55 upon absorbing
u b = gρd 2b 18μ (76)
a small amount of SO2 at the thirtieth minute
approximately and the outlet SO2 concentration
where the mean diameter of the bubbles:
increased more and more quickly from 0. This
indicated that NH4HCO3 in the liquid phase was
d b = 0.18d1/2 1/3
0 R e0 (77)
consumed completely and the absorption process
changed from a fast chemical reaction to a weak one.
R e0 = 4QρG πd 0μG and Re0<2000 (78) Thus, the absorption rate becomes smaller and the
outlet concentration gradually increases.
Gas holdup (Meikap et al., 2002):
2500

( ) ( gd )
18 1 12
εg = 0.25 d 2R ρ L g δ L 3 2
Rρ L μ 2L
-3
Inlet SO2 concentration g·m

2.6785
(79) 2000 5.3571
-3

6.6964
ub gd R
Oulet SO2 concentration /mg·m

8.0357

1500
Interfacial area per unit volume (Meikap et al.,
2002):
1000

a = 48.7 ( u b μ L )
0.51
(80)
500

The reactor height was discretized in a spatially

uniform grid and the finite difference method was 0
applied. These coupled partial differential equations 0 10 20 30 40 50

were solved numerically by using Matlab. Time /min

Figure 2: Effect of the inlet SO2 concentrations on

RESULTS AND DISCUSSIONS
In the bubble column model, the liquid-side mass
transfer coefficient for the absorption of SO2 into
aqueous NH4HCO3 solution was calculated from
Eqs. (57) and (58). The comparisons between the
calculated results and the experimental results for the
fast reactive absorption of SO2 into aqueous
NH4HCO3 solution in the bubble column under
different operating conditions were made and are
shown in Fig. 2-4. It can be seen that the calculated
results agree well with the experimental results,
which verifies that the mass transfer model based on
the aforementioned scheme is reasonable and
accurate.
the outlet SO2 concentrations (concentration of
ammonium 0.05%, gas flow rate 20 L·min-1). Scatter
plots for experimental results and smooth curves for
the modeling results.
Effect of the Concentration of NH4HCO3
The effect of the concentration of NH4HCO3 on
the gas absorption was presented in Fig. 4. From Fig.
4, the duration time for complete SO2 removal
increased with the increase of the concentration of
NH4HCO3. This is because when the concentration
of NH4HCO3 is high, more SO2 must be introduced
into the system to react completely with NH4HCO3.
Thus, the duration time for complete SO2 removal

Brazilian Journal of Chemical Engineering

 Mass Transfer of SO2 Absorption with an Instantaneous Chemical Reaction in a Bubble Column 559

increased with the increase of the concentration of concentrations increased more significantly as the
NH4HCO3. gas flow rate increased. According to Eq. (76),
increasing the gas flow rate makes the bubbles’
9
rising velocity higher, which contributes to a very
8
2000 high fractional rate of surface renewal, liquid-side

-3
mass transfer coefficient, gas holdup and interfacial

Oulet SO2 concentration /mg¡¤m

7 1500
area. Thus, NH4HCO3 is consumed quickly and the
duration time for complete SO2 removal becomes
6 shorter. In the last period of the SO2 absorption, the
1000
absorption process is actually water absorbing SO2,
pH

5
and therefore the higher the gas flow rate, the faster
4 500 the concentration of SO2 in the liquid phase increases.
PH
oulet SO2 concentration
Thus, the outlet SO2 concentrations increase more and
3 more quickly with the increase of the gas flow rate.
0

2
0 10 20 30 40 50 2500
-1
Time /min Gas flow rate L·min

20
2000 30
Figure 3: Variation of the outlet SO2 concentrations

-3
60

Oulet SO2 concentration /mg·m

and pH value varying with time (concentration of the
inlet SO2 was 2.6785 g·m-3, concentration of ammo- 1500

nium bicarbonate 0.05%, gas flow rate 20 L·min-1).

Scatter plots for experimental results and smooth 1000
curves for the modeling results.
500

0
2000
Concentration of NH4HCO3
-3

0 10 20 30 40 50 60
Oulet SO2 concentration /mg¡¤m

0.20%
0.15% Time /min
1500 0.10%
0.05%

Figure 5: Effect of the gas flow rate on the outlet

1000 SO2 concentrations (inlet concentration of SO2 was
5.3571 g·m-3, concentration of ammonium bicarbonate
0.05%. Scatter plots for experimental results and
500
smooth curves for the modeling results.

0
0 10 20 30 40 50 60 CONCLUSION
Time /min

Gas absorption accompanied by an instantaneous

Figure 4: Effect of concentrations of ammonium irreversible chemical reaction in bubble columns has
bicarbonate solution on the outlet SO2 concentrations been studied theoretically and experimentally. A
(inlet concentration of SO2 was 2.6785 g·m-3, gas model based on the surface-renewal model and
flow rate 20 L·min-1). Scatter plots for experimental penetration theory has been developed and analyti-
results and smooth curves for the modeling results. cally solved. The analytical expressions for the time-
average mass transfer coefficient and the enhance-
Effect of the Gas Flow Rate ment factor have been derived. The time-average
mass transfer coefficient and the enhancement factor
Figure 5 presents the effect of the gas flow rate are functions of the saturated concentration of the
on the gas absorption. It can be seen from Figure 5 gas-phase reactant in the interface, the concentration
that the duration time for complete SO2 removal of the liquid-phase reactant in the bulk liquid, the
decreased rapidly with increasing gas flow rate. In diffusion coefficients of the gas-phase and liquid-
the last period of the SO2 absorption, the outlet SO2 phase reactants in the liquid and the stoichiometric

Brazilian Journal of Chemical Engineering Vol. 30, No. 03, pp. 551 - 562, July - September, 2013
560 Xiaolei Li, Chunying Zhu, Sumin Lu and Youguang Ma

coefficient of the reaction. The time-average mass DG diffusivity in the gas phase m2·s-1
transfer coefficient is correlated with the fractional DSO2 diffusivity of SO2 in the m2·s-1
rate of surface renewal and influenced by the bubble liquid phase
rising velocity. The fast reactive absorption of SO2 D diffusivity of HSO3− in the m2·s-1
HSO3−
into aqueous NH4HCO3 solutions with different liquid phase
concentrations in a bubble column was experimen- m2·s-1
D
HCO3− diffusivity of HCO3− in the
tally investigated. An absorption model accompanied
by fast chemical reaction has been developed and liquid phase
numerically solved. The prediction results agree well dR column diameter m
with the experimental results, which validates the g gravitational acceleration m·s-2
mass transfer model proposed. G defined by Eq. (19)
h defined by Eq. (25)
H average distance of m
ACKNOWLEDGMENTS approach of eddies to the
interface
We gratefully acknowledge the financial support JA molar flux of diffusion of mol·m-2·s-1
for this project from the National Natural Science species A relative to the
Foundation of China (No 21076139), and the molar average velocity
Program of Introducing Talents of Discipline to kG gas side mass transfer m·s-1
Universities (Grant No B06006). coefficient
kL fluid-film mass transfer m·s-1
coefficient
NOMENCLATURE Ki reaction rate constant (i=1,
2, 3, 4, 5, 6)
a specific interfacial surface m2·m-3 KL time-average mass transfer m·s-1
A exposed area of time- m2 coefficient in the liquid
dependent surface phase
A0 initial area of the surface m2 KL,tol total mass transfer m·s-1
exposed to gas coefficient of SO2
total area of the surface m2 l length scale of eddies in the m
A∞ viscous dissipation range
exposed to gas
A* saturation concentration of A mol·L-1 NA overall flux of species A mol·m-2·s-1
B* saturation concentration of B mol·L-1 relative to a phase boundary
C molar concentration at the mol·L-1 R defined by Eq. (72)
interface Re0 the orifice Reynolds number
cA molar concentration of mol·L-1 S fractional rate of surface s-1
species A renewal
cB molar concentration of mol·L-1 T dimensionless time variable
species B defined by Eq. (13)
CA dimensionless concentration ug superficial gas velocity m·s-1
profile defined by Eq. (14) ul superficial liquid velocity m·s-1
CB dimensionless concentration x rectangular coordinate m
profile defined by Eq. (15) y rectangular coordinate m
Ci constant z rectangular coordinate m
CL concentration of the total mol·L-1 Z1 dimensionless position
sulfur in the liquid variable defined by Eq. (11)
CSO2, L concentration of SO2 in the mol·L-1 Z2 dimensionless position
liquid variable defined by Eq. (12)
DA diffusivity of A m2·s-1
DB diffusivity of B m2·s-1 Greek Letters
Dg dispersion coefficient, gas m2·s-1
phase α defined by Eq. (22)
DL dispersion coefficient, liquid 2 -1
m ·s ε energy dissipation rate per W·kg-1
phase unit mass

Brazilian Journal of Chemical Engineering

 Mass Transfer of SO2 Absorption with an Instantaneous Chemical Reaction in a Bubble Column 561

εG gas hold-up Dankwerts, P. V., Significance of liquid-film coeffi-

θ time s cients in gas absorption. Ind. Eng. Chem., 43,
θ0 characteristic constant s 1460-1467 (1951).
defined for bubble column de Lind van Wijngaarden, G., Versteeg, G. F.,
system Beenackers, A. A. C. M., Mass-transfer enhance-
κ kinematic viscosity m2·s-2 ment factors for reversible gas–liquid reactions:
μG viscosity of gas phase kg·m-1·s-1 Comparison of DeCoursey’s and Onda’s methods.
μL viscosity of liquid phase kg·m-1·s-1 Chem. Eng. Sci., 41, 2440-2442 (1986).
ρG gas density kg·m-3 Ebrahimi, S., Picioreanu, C., Kleerebezem, R.,
ρL liquid density kg·m-3 Heijnen, J. J., van Loosdrecht, M. C. M., Rate-
σ distance from the interface m based modelling of SO2 absorption into aqueous
to the bulk of the liquid NaHCO3/Na2CO3 solutions accompanied by the
σ1 dimensionless distance from desorption of CO2. Chem. Eng. Sci., 58, 3589-
the interface to the bulk of 3600 (2003).
the liquid Garg, R., Nair, S., Bhaskarwar, A. N., Mass transfer
σL surface tension of the liquid N·m-1 with instantaneous chemical reaction in finite
υ molar average velocity; m·s-1 gas-liquid systems. Chem. Eng. J., 76, 89-98
velocity scale of eddies in (2000).
viscous dissipation range Harriott, P., A random eddy modification of the
Φ1 defined by Eq. (19) penetration theory. Chem. Eng. Sci., 17, 149-154
Φ2 defined by Eq. (19) (1962).
Jajuee, B., Margaritis, A., Karamanev, D., Bergougnou,
Subscripts M. A., Application of surface- renewal-stretch
model for interface mass transfer. Chem. Eng.
A gas phase reactant Sci., 61, 3917-3929 (2006).
B liquid phase reactant Kakaraniya, S., Kari, C., Verma, R., Mehra, A., Gas
absorption in slurries of fine particles: SO2-
Mg(HO)2-MgSO3 system. Ind. Eng. Chem. Res.,
REFERENCES 46, 1904-1913 (2007).
Last, W., Stichlmair, J., Determination of mass transfer
Angelo, J. B., Lightfoot, E. N., Howard, D. W., parameters by means of chemical absorption.
Generalization of the penetration theory for Chem. Eng. Technol., 25, No. 4, 385-391 (2002).
surface stretch: Application to forming and oscil- Martínez, I., Casas, P. A., Simple model for CO2
lation drops. AIChE J., 12, 751-760 (1966). absorption in a bubbling water column. Braz. J.
Bokotko, R. P., Hupka, J., Miller, J. D., Flue gas Chem. Eng., 29, 107-111 (2012).
treatment for SO2 removal with air-sparged Mehra, A., Gas absorption in slurries of finite-
hydrocyclone technology. Environ. Sci. Technol., capacity microphases. Chem. Eng. Sci., 45, 1525-
39, 1184-1189 (2005). 1538 (1990).
Cho, J. S., Wakao, N., Determination of liquid-side Meikap, B. C., Kundu, G., Biswas, M. N., Modeling
and gas-side volumetric mass transfer coefficients of a novel multi-stage bubble column scrubber for
in a bubble column. J. Chem. Eng. Jpn., 21, 576- flue gas desulfurization. Chem. Eng. J., 86, 331-
581 (1988). 342 (2002).
da Silva, E. A. B., Souza, D. P., de Souza, A. A. U., Meldon, J. H., Olawoyin, O. O., Bonanno, D.,
de Souza, S. M. A. G. U., Prediction of effective Analysis of mass transfer with reversible
diffusivity tensors for bulk diffusion with chemi- chemical reaction. Ind. Eng. Chem. Res., 46,
cal reactions in porous media. Braz. J. Chem. 6140-6146 (2007).
Eng., 24, 47-60 (2007). Moo-Young, M., Blanch, H. W., Design of bio-
Danish, M., Sharma, R. K., Ali, S., Gas absorption chemical reactors mass transfer criteria for simple
with first order chemical reaction in a laminar and complex systems. Adv. Biochem. Eng./
falling film over a reacting solid wall. Appl. Biotechnol., 19, 1-69 (1981).
Math. Model., 32, 901-929 (2008). Wang, T., Wang, J., Jin, Y., Theoretical prediction of
Danckwerts, P. V., Unsteady-state diffusion or heat- flow regime transition in bubble columns by the
conduction with moving boundary. Trans. population balance mode. Chem. Eng. Sci., 60,
Faraday Soc., 46, 701-712 (1950). 6199-6209 (2005).
Brazilian Journal of Chemical Engineering Vol. 30, No. 03, pp. 551 - 562, July - September, 2013
562 Xiaolei Li, Chunying Zhu, Sumin Lu and Youguang Ma

Yue, J., Rebrov, E. V., Schouten, J. C., Enhancement Zhang, Y., Zheng, L., Analysis of MHD thermo-
factor for gas absorption in a finite liquid layer. Part solutal Marangoni convection with the heat
1: Instantaneous reaction in a liquid in plug flow. generation and a first-order chemical reaction.
Chem. Eng. Technol., 35, No. 4, 679-692 (2012). Chem. Eng. Sci., 69, 449-455 (2012).

Brazilian Journal of Chemical Engineering